Article Text

Original research
IL-20 subfamily cytokines impair the oesophageal epithelial barrier by diminishing filaggrin in eosinophilic oesophagitis
  1. Tanay Kaymak1,
  2. Berna Kaya1,
  3. Philipp Wuggenig1,
  4. Sandro Nuciforo2,
  5. Andreas Göldi3,
  6. Swiss EoE Cohort Study Group (SEECS),
  7. Franz Oswald4,
  8. Julien Roux1,5,
  9. Mario Noti6,
  10. Hassan Melhem1,
  11. Petr Hruz3,
  12. Jan Hendrik Niess1,3,7
    1. 1Department of Biomedicine, Gastroenterology, University of Basel, Basel, Switzerland
    2. 2Department of Biomedicine, Hepatology, University of Basel, Basel, Switzerland
    3. 3Department of Gastroenterology, Clarunis - University Center for Gastrointestinal and Liver Diseases, Basel, Switzerland
    4. 4Department of Internal Medicine I, University Hospital Ulm, Ulm, Germany
    5. 5Swiss Institute of Bioinformatics, Basel, Switzerland
    6. 6Institute of Pathology, University of Bern, Bern, Switzerland, Current address: Nestlé SA, Nestlé Research, Nestlé Institute of Health Sciences, Department of Gastrointestinal Health Immunology, Vers-Chez-les-Blancs, Lausanne, Switzerland
    7. 7Department of Clinical Research, University of Basel, Basel, Switzerland
    1. Correspondence to Dr Jan Hendrik Niess, Department of Biomedicine, Gastroenterology, University of Basel, Basel, Switzerland; janhendrik.niess{at}unibas.ch

    Abstract

    Objective Disruption of the epithelial barrier plays an essential role in developing eosinophilic oesophagitis (EoE), a disease defined by type 2 helper T cell (Th2)-mediated food-associated and aeroallergen-associated chronic inflammation. Although an increased expression of interleukin (IL)-20 subfamily members, IL-19, IL-20 and IL-24, in Th2-mediated diseases has been reported, their function in EoE remains unknown.

    Design Combining transcriptomic, proteomic and functional analyses, we studied the importance of the IL-20 subfamily for EoE using patient-derived oesophageal three-dimensional models and an EoE mouse model.

    Results Patients with active EoE have increased expression of IL-20 subfamily cytokines in the oesophagus and serum. In patient-derived oesophageal organoids stimulated with IL-20 cytokines, RNA sequencing and mass spectrometry revealed a downregulation of genes and proteins forming the cornified envelope, including filaggrins. On the contrary, abrogation of IL-20 subfamily signalling in Il20R2−/− animals resulted in attenuated experimental EoE reflected by reduced eosinophil infiltration, lower Th2 cytokine expression and preserved expression of filaggrins in the oesophagus. Mechanistically, these observations were mediated by the mitogen-activated protein kinase (MAPK); extracellular-signal regulated kinases (ERK)1/2) pathway. Its blockade prevented epithelial barrier impairment in patient-derived air–liquid interface cultures stimulated with IL-20 cytokines and attenuated experimental EoE in mice.

    Conclusion Our findings reveal a previously unknown regulatory role of the IL-20 subfamily for oesophageal barrier function in the context of EoE. We propose that aberrant IL-20 subfamily signalling disturbs the oesophageal epithelial barrier integrity and promotes EoE development. Our study suggests that specific targeting of the IL-20 subfamily signalling pathway may present a novel strategy for the treatment of EoE.

    • barrier function
    • oesophagitis
    • cytokines
    • immunology

    Data availability statement

    Data are available in a public, open access repository. Data are available upon reasonable request. Bulk RNA sequencing (GEO accession GSE181261; https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE181261), single-cell RNA-seq (GEO accession GSE190482; https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE190482) and proteomics (PXD031509; project DOI: 10.6019/PXD031509) datasets are publicly available in open access repositories.

    http://creativecommons.org/licenses/by-nc/4.0/

    This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.

    Statistics from Altmetric.com

    WHAT IS ALREADY KNOWN ON THIS TOPIC

    • Eosinophilic oesophagitis (EoE) is a type 2 helper T cell (Th2)-mediated food-associated and aeroallergen-associated chronic inflammatory disease with increasing prevalence and puzzling pathophysiology.

    • Th2 cytokines interleukin (IL)-4 and IL-13 negatively impact oesophageal barrier integrity, resulting in increased permeability.

    • Members of the IL-20 cytokine family mediate crosstalk between immune cells and epithelial cells.

    WHAT THIS STUDY ADDS

    • Active EoE is associated with increased expression of IL-20 subfamily cytokines in the oesophagus and serum.

    • IL-20 cytokines lower the expression of genes involved in oesophageal barrier integrity, including filaggrins.

    • Abrogation of IL-20 cytokine signalling attenuates experimental EoE in mice.

    • Inhibition of STAT3 signalling reinforces IL-20 subfamily-mediated reduction of filaggrins and aggravates experimental EoE in mice.

    • Pharmacological inhibition of ERK rescues IL-20 subfamily-mediated epithelial barrier impairment and attenuates experimental EoE in mice.

    HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE AND/OR POLICY

    • Targeting aberrant IL-20 subfamily signalling may represent a novel therapeutic approach for the treatment of EoE.

    • Ongoing clinical development of ERK inhibitors for several disease entities reflects their therapeutic potential. In EoE, ERK inhibitors could be used to restore epithelial integrity.

    Introduction

    Eosinophilic oesophagitis (EoE) is a food allergen-driven chronic inflammatory disease characterised by solid food dysphagia and food impaction in adults and food refusal and failure to thrive in children.1 2 Inflammation in EoE is predominated by a type 2 immune response triggered by food antigens.3 4 However, the exact immunological pathways that lead to EoE are still poorly understood. An increasing number of studies illustrate that cytokines such as type 2 helper T cell (Th2)-derived interleukin (IL)-13 elicit transcriptional changes in epithelial cells, altering the integrity of the oesophageal epithelium.5–7

    Emerging evidence suggests impaired epithelial barrier integrity is a dominant element in EoE pathophysiology. Genome-wide association studies have identified most associated polymorphisms in the oesophageal epithelium, including variants in the filament aggregating protein filaggrin.8–10 Coinciding with this, reduction of proteins, such as filaggrins, desmogleins, occludin and claudins involved in the formation of the apical junction complex is observed in EoE.7 11 12 Epithelial barrier impairment by filaggrin reduction results from disturbed protease–protease inhibitor homeostasis and aberrant cytokine signalling in the oesophagus.8 13 14

    Growing evidence indicates that the IL-10 cytokine family significantly contributes to the crosstalk between the immune system and epithelial cells. Among these is the IL-20 subfamily, comprising IL-19, IL-20 and IL-24.15 All IL-20 subfamily members share the type 1 IL-20 receptor formed by IL-20Rβ and IL-20Rα subunits. On the other hand, IL-20 and IL-24 but not IL-19 can also signal through the type 2 IL-20 receptor created by IL-20Rβ and IL-22Rα1.16 The activation of both type 1 and type 2 IL-20 receptors results in STAT3 phosphorylation,16 suggesting overlapping functions of IL-19, IL-20 and IL-24 by receptor sharing and conserved signalling cascades. IL-20 cytokine family members target keratinocytes17 in the skin, where all three cytokines can induce acanthosis and S100A7 expression.17 The IL-20 subfamily has been linked to Th2-mediated allergic skin inflammation and allergen-induced airway inflammation.18–20 In patients with atopic dermatitis, IL-24 downregulates the expression of filaggrin in keratinocytes.18 To the same effect, IL-19 and IL-24 mediate epidermal hyperplasia induced by intradermal IL-23 injection in an IL-20Rβ -dependent manner.21 Moreover, transgenic animals overexpressing IL-20 or IL-24 developed epidermal hyperplasia and abnormalities in keratinocyte differentiation,15 22 consolidating the relationship between IL-20 subfamily cytokines and epithelial cells. However, it has not yet been studied if the aberrant activity of the IL-20 subfamily affects keratinocyte differentiation in the oesophagus. Thus, we explored whether dysregulation of the IL-20 subfamily could impair the oesophageal epithelial barrier during EoE.

    Materials and methods

    The resources used are listed in online supplemental tables 4–6. Further details of resources and methods are described in the online supplemental data file.

    Supplemental material

    Oesophageal biopsies from human subjects

    Patients referred for diagnostic oesophagogastroduodenoscopy were evaluated for EoE or eligibility as control. Biopsies were taken from the distal and proximal oesophagus and assessed by an independent pathologist. Control subjects were defined as having no endoscopic and histological signs of oesophagitis, nil eosinophils per high-power field (HPF), no history of EoE and no hints of motility disorders.

    Patients with EoE were defined as having a confirmed EoE diagnosis or presenting with ≥15 eosinophils per HPF in proximal biopsies, symptoms suggestive of oesophageal dysfunction, such as solid food dysphagia or food impaction, and not having any endoscopic and histological signs of GERD.

    Detailed patient characteristics are listed in online supplemental table 1.

    Patient and public involvement statement

    Patients or the public were not involved in our study’s design, conduct, reporting or dissemination plans.

    Statistical analysis

    Data are shown as dot plots and display individual values with medians. GraphPad Prism software was used to generate the graphs and perform statistical analysis. Depending on the experimental setting, p values were calculated using either Mann-Whitney U, Wilcoxon, Kruskal-Wallis or two-way analysis of variance tests. The Grubbs test further analysed the data to identify outliers. P values were shown as follows: *p≤0.05, **p≤0.01, ***p≤0. 001 and ****p≤0. 0001.

    Results

    Elevated IL-20 subfamily cytokines in EoE and type 1 IL-20 receptor expression by the oesophageal epithelium

    Recent reports have suggested an association of the IL-20 subfamily with Th2-mediated allergies and allergen-induced airway inflammation,18–20 but as of yet, neither their function in the oesophagus nor a potential role in the context of EoE has been studied. Patients with active EoE have significantly increased IL19 and IL20 but not IL24 expression in oesophageal biopsies compared with control individuals (figure 1A), while topical corticosteroid treatment in patients with active EoE reduced IL19 and IL20 expression in oesophageal biopsies (figure 1A,B). Serum concentrations of all IL-20 subfamily cytokines, including IL-24, were increased in patients with active EoE compared with control individuals, while treatment with topical corticosteroids lowered IL-20 subfamily cytokines in the serum of patients achieving remission (figure 1C), suggesting for the first time that IL-20 subfamily cytokines may play a role in the active phase of EoE.

    Figure 1

    Increased IL-20 cytokines in active EoE. (A) IL19, IL20, IL24 mRNA expression levels in the oesophagus of control individuals, active EoE and inactive EoE treated with topical corticosteroids measured by RT-qPCR. (B) Paired IL19, IL20 and IL24 mRNA expression levels in the oesophagus measured by RT-qPCR before and during treatment with topical corticosteroids. (C) IL-19, IL-20 and IL-24 serum concentrations of indicated groups measured by ELISA. (D) IL20Rα, IL20Rβ and IL22Rα1 mRNA expression in control individuals, active EoE and inactive EoE treated with topical corticosteroids by RT-qPCR. (E) Representative control staining, IL-20Rα and IL-20Rβ staining of the proximal oesophagus obtained from control individuals and patients with EoE. White arrowheads indicate eosinophils; black arrowheads indicate IL-20Rα and IL-20Rβ staining. Scale bars, 50 µm. Each dot or line represents one individual with medians. *P≤0.05, **P≤0.01 by Mann-Whitney U test. EoE, eosinophilic oesophagitis; GAPDH, glyceraldehyde 3-phosphate dehydrogenase; IL, interleukin; RT-qPCR, reverse transcription-quantitative polymerase chain reaction.

    To identify cellular targets of IL-20 subfamily cytokines, we performed a RT-qPCR analysis of the IL-20R subunits. The expression analysis indicated IL20Rα, IL20Rβ and IL22Rα1 expressions in oesophageal biopsies with lower IL20Rα and IL20Rβ expressions in patients with active EoE and topical corticosteroid-treated inactive EoE compared with control individuals (figure 1D). Immunohistochemistry-stained sections from the oesophagus confirmed the predominant expression of IL-20Rα and IL-20Rβ by oesophageal epithelial cells (figure 1E). Collectively, we demonstrate dysregulated IL-20 subfamily cytokine responses in patients with active EoE and expression of the type 1 IL-20 receptor by the oesophageal epithelium.

    IL-20 cytokines reduce filaggrin expression in patient-derived oesophageal organoids

    To directly address the impact of the IL-20 subfamily on the oesophageal epithelium, we have adopted a protocol to generate oesophageal organoids from patient-derived biopsies.23 Cells isolated from oesophageal biopsies of control individuals and patients with EoE form self-organising organoids, which proliferate without morphological differences over 11 days (online supplemental figure 1). IL-20Rα expression increased during the culture period with the highest expression on day 11, while IL-20Rβ is constitutively expressed throughout the culture period in controls and EoE (online supplemental figure 1).

    We stimulated organoids from control individuals with IL-19, IL-20 and IL-24 to elucidate the effect of the IL-20 subfamily on the transcriptome and proteome of the oesophageal epithelium.24 25 Using a principal component analysis (PCA), we observed that interindividual differences dominated the RNA sequencing (RNA-seq) of paired non-stimulated and stimulated organoids. Deeper principal components revealed segregation of non-stimulated from stimulated oesophageal organoids (figure 2A). To gain insights into molecular mechanisms by which IL-19, IL-20 and IL-24 affect the oesophageal squamous epithelium, we performed a differential expression analysis between stimulated and non-stimulated oesophageal organoids followed by gene set enrichment analysis (GSEA). Differentially expressed genes included notably keratins (ie, KRT4, KRT13 and KRT24), components of desmosomes (ie, DSG1 and 4, DSC1 and 2 and DSP), members of the tight junction complex (ie, OCLN and CLDN17), filament-associated proteins, such as FLG and FLG2, and members of the serine protease inhibitor kazal (SPINK)-type family, such as SPINK7 (figure 2B,C).

    Figure 2

    The stimulation of patient-derived oesophageal organoids with IL-19, IL-20 and IL-24 lowers the expression of genes involved in barrier integrity. (A) PCA based on the normalised expression levels in the RNA-seq samples of paired non-stimulated and IL-20 subfamily (IL-19, IL-20 and IL-24) stimulated oesophageal organoids from control individuals (numbered 12, 14, 15, 17 and 18). (B) Volcano plot showing the results of the differential expression analysis between non-stimulated and stimulated human oesophageal organoids. Only genes with a p value lower than 10−7 are labelled, and genes of the highest interest for our study are highlighted in bold. (C) Heatmap showing the centred and scaled expression levels of a selected subset of significant differentially expressed genes associated with the epithelium. Genes of the highest interest for our study are highlighted in bold. (D) GSEA results based on differentially expression results from the RNA-seq dataset. The x-axis depicts the average absolute log-fold change across genes within each category. Only categories with a p value lower than 10−7 and absolute average log-fold change of >0.15 are labelled. (E) PCA based on normalised protein levels in the proteomics samples of paired non-stimulated and IL-20 subfamily (IL-19, IL-20 and IL-24) stimulated oesophageal organoids from control individuals (numbered 38, 39, 44, 49 and 50). (F) Volcano plot showing the differential protein expression analysis results between non-stimulated and stimulated human oesophageal organoids. All significant proteins are labelled. (G) Heatmap showing the centred and scaled expression levels of a subset of proteins from the genes used (C) and present in the proteomics dataset. Proteins of the highest interest for our study are highlighted in bold. (H) Results of the GSEA based on differentially expression results from the proteomics dataset. GSEA, gene set enrichment analysis; IL, interleukin; PCA, principal component analysis; RNA-seq, RNA sequencing.

    Furthermore, we noted an increased expression of the IL-13 receptor subunits IL-4Ra, IL-13Ra1 and IL-13Ra2 (online supplemental figure 2). Above all, downregulated genes were associated with keratinisation, cornification, cornified envelope and epidermal differentiation (figure 2D).24

    We compared these results to the transcriptome differences previously reported in esophageal keratinocytes stimulated with IL-13 (GSE65335),26 SPINK7-deficient esophageal keratinocytes (GSE103356)27 and the EoE transcriptome (GSE58640)28 (online supplemental figure 3). The gene ontology (GO) terms cornification, cornified envelope, epidermal cell differentiation and keratinisation were downregulated in all four datasets. Interestingly, the GO term keratin filament was only downregulated in our dataset and the EoE transcriptome. Additional IL-20 subfamily-specific effects on the epithelium were reflected by downregulation of the GO terms tight junction, desmosome, epidermis morphogenesis and apical junctional complex in our but not in the other reported transcriptomic datasets (online supplemental figure 3).

    Consistent with the transcriptomic data, proteomics analysis of paired non-stimulated and stimulated oesophageal organoids from different control individuals displayed high interindividual differences in the PCA. Similarly, segregation of non-stimulated oesophageal organoids from their stimulated counterparts was seen in deeper principal components (figure 2E). Differential expression analysis and GSEA on the protein levels revealed that the extent of changes associated with IL-20 subfamily cytokine stimulation was lower than that observed at the transcriptome level. However, similarly, proteins related to epithelial cornification and differentiation were downregulated (figure 2F–H).25

    Despite the altered expression of genes and proteins involved in epithelial cornification and differentiation, we did not observe morphological changes in the organoids on cytokine stimulation (figure 3A). However, RT-qPCR and immunohistochemistry confirmed that stimulating oesophageal organoids from control individuals and patients with EoE with IL-19, IL-20 and IL-24 reduced FLG, FLG2 and SPINK7 expressions (figure 3B–D), while the receptor expression remained unchanged (figure 3E). Moreover, in patients with active EoE, we observed reduced FLG, FLG2 and SPINK7 expressions in the oesophagus compared with controls and patients with inactive EoE treated with topical corticosteroids (figure 3F). Immunohistochemistry confirmed reduced FLG and FLG2 levels in the oesophageal epithelium of patients with active EoE, which was partially restored in inactive EoE (figure 3G,H). Together, our data illustrate that the stimulation of patient-derived oesophageal organoids with IL-19, IL-20 and IL-24 reduced the expression of genes and proteins involved in maintaining the epithelial barrier, including the filaggrin family and the serine protease inhibitor SPINK7.

    Figure 3

    Stimulation with IL-19, IL-20 and IL-24 decreases filaggrin expression. (A) Bright-field images of oesophageal organoids from control individuals at the indicated time. On day 11, organoids were stimulated with IL-19, IL-20 and IL-24 (IL-20s). Scale bars, 50 µm. (B) FLG, FLG2 and SPINK7 mRNA expression of paired non-stimulated (NS) and with IL-19, IL-20 and IL-24 (IL-20s) stimulated organoids from control individuals or (C) patients with EoE measured by RT-qPCR. (D) Immunohistochemistry for FLG or FLG2 and control of non-stimulated (NS) and with IL-19, IL-20 and IL-24 (IL-20s) stimulated oesophageal organoids from control individuals. Scale bars, 50 µm. (E) IL20Rα, IL20Rβ and IL22Rα1 mRNA expression in non-stimulated (NS) and with IL-19, IL-20 and IL-24 (IL-20s) stimulated oesophageal organoids. (F) FLG, FLG2 and SPINK7 mRNA expression in the proximal oesophagus of control individuals, active EoE and inactive EoE. (G) FLG and FLG2 staining with haematoxylin counterstaining in the proximal oesophagus from control individuals, active EoE and inactive EoE. Scale bars, 50 µm. (H) Quantification of FLG and FLG2 staining using QuPath software. Data are presented as individual values with medians, with each dot or line representing one biological replicate. *P≤0.05, **P≤0.01, ***P≤0.001 by Mann-Whitney U or Wilcoxon test. EoE, eosinophilic oesophagitis; IL, interleukin; NS, non-stimulated; RT-qPCR, RT-qPCR, reverse transcription-quantitative polymerase chain reaction.

    Increased IL-20 cytokine expression in murine EoE-like disease model

    Because sensitisation to food particles in patients with EoE can occur via the skin,29–31 we decided to adopt an experimental EoE mouse model induced by epicutaneous allergen sensitisation (online supplemental figure 4A).32 Ovalbumin (OVA)-skin sensitised and orally challenged (sens+chal) mice developed experimental EoE characterised by oesophageal eosinophil infiltration (online supplemental figure 4B–E). Untreated (ctrl), non-sensitised (non-sens) and sensitised but not challenged (non-chal) animals did not develop experimental EoE (online supplemental figure 4B–E). Consistent with findings in human patients with EoE, mice with experimental EoE demonstrated significantly increased Il19 and increased Il20 and Il24 expression levels in the oesophagus compared with ctrl and non-chal animals (figure 4A). Flow cytometry analysis of Il19-tdTomato (Il19tdT) reporter animals with experimental EoE indicated that macrophages in the oesophagus of mice with EoE produce Il19 (online supplemental figure 4F,G).

    Figure 4

    Abrogation of IL-20 cytokine signalling attenuates experimental EoE. (A) Il19, Il20 and Il24 mRNA expressions in the murine oesophagus by RT-qPCR. (B) H&E staining of oesophageal sections of WT, Il19tdT and Il20R2−/− animals. Black arrowheads indicate eosinophils. Scale bars, 50 µm. (C) Quantification of eosinophils (per HPF) from (B). (D) Frequencies and absolute numbers of eosinophil infiltrates in the oesophagus as assessed by flow cytometry. (E) Representative flow cytometry plots of eosinophil frequencies in the oesophagus. Data are presented as individual values with medians, with each dot representing one biological replicate. *P<0.05, **P<0.01, ****P<0.0001, by Mann-Whitney U test. ctrl, non-sensitised+non-challenged; HPF, high-power field; IL, interleukin; non-chal, sensitised+non-challenged; non-sens: non-sensitised+challenged; sens+chal: sensitised+challenged; RT-qPCR, reverse transcription-quantitative polymerase chain reaction; WT, wild type.

    In the GI tract of mice, the expression of type 1 IL-20 receptor subunits Il20rα and Il20rβ was highest in the oesophagus and forestomach (squamous epithelium) compared with lower expression in the glandular stomach and the small intestine (columnar epithelium) (online supplemental figure 4H). Noteworthily, murine bone marrow-derived macrophages (BMDMs) only express Il20rβ but not Il20rα and Il22ra1 (online supplemental figure 4I), suggesting that macrophages do not have a functional IL-20 receptor. Consistently, single-cell RNA-seq of CD45+ immune cells from mice with experimental EoE indicates that immune cells generally do not express a functional IL-20 receptor (online supplemental figure 4J).33 In summary and consistent with an observed phenotype in humans, experimental EoE in mice is characterised by oesophageal eosinophilia and increased Il19, Il20 and Il24 expressions in the oesophagus.

    IL-20R deficiency is protective in the EoE mouse model

    Next, we induced the experimental EoE mouse model in wild type (WT), IL-19-deficient (Il19tdT) and Il20R2−/− mice, in which signalling of all three IL-20 cytokines is impaired. Detection of total IgE and OVA-specific IgE by ELISA confirmed successful sensitisation in WT, Il19tdT and Il20R2−/− mice (online supplemental figure 5A,B). Chal+sens Il20R2−/− mice had significantly reduced eosinophils in the oesophagus compared with WT and Il19tdT mice, evidenced by H&E staining (figure 4B,C) and flow cytometry (figure 4D,E) at day 4 of OVA challenge when WT animals have developed severe experimental EoE. While Il19tdT mice had reduced eosinophil numbers in the oesophagus compared with WT mice, specific deletion of Il19 from CX3CR1+ macrophages in a novel conditional Il19-deficient mouse line (online supplemental figure 5C) suggested that macrophage-derived IL-19 alone does not account for EoE in mice (online supplemental figure 5D–G). Increased oesophageal eosinophilia in chal+sens WT animals was associated with elevated Il5 and Il13 expression in the oesophagus (online supplemental figure 5H,I). At the same time, EoE-protected Il20R2−/− mice had lower Il5 and Il13 expression than WT mice (online supplemental figure 5H,I). Intriguingly, non-chal and sens+chal WT mice had reduced Flg and Flg2 but not Spink7 gene expression in the oesophagus compared with ctrl and non-sens animals (figure 5A). Both sensitisation with MC903 and OVA challenge in Il19tdT and Il20R2−/− mice did not reduce Flg and Flg2 gene expression (figure 5B,C). Immunohistochemistry showed reduced suprabasal Flg2 in the stratum spinosum of the oesophagus with retained expression in the stratum corneum of non-chal and sens+chal WT and Il19tdT mice, whereas Flg2 expression was preserved in all oesophageal epithelial layers of Il20R2−/− mice (figure 5D,E). In summary, EoE is attenuated in Il20R2−/− mice which may be linked to preserved expression of the filaggrin family in the oesophagus.

    Figure 5

    Abrogation of IL-20 signalling preserves filaggrin expression in experimental EoE. (A–C) Flg, Flg2 and Spink7 mRNA expression in (A) WT, (B) Il19tdT and (C) Il20R2−/− mice quantified by RT-qPCR. (D+E) Flg2 immunohistochemistry of WT, Il19tdT and Il20R2−/− animals; (D) percentage of Flg2 stained area (analysed with Fiji (ImageJ V.2.0.0-rc-68/1.52 hour)) and (E) representative images. Scale bars, 50 µm. Data are presented as individual values with medians, with each dot representing one biological replicate. *P<0.05, by Mann-Whitney U test. ctrl, non-sensitised+non-challenged; EoE, eosinophilic oesophagitis; IL, interleukin; non-chal, sensitised+non-challenged; non-sens, non-sensitised+challenged; sens+chal, sensitised+challenged; RT-qPCR, reverse transcription-quantitative polymerase chain reaction; WT, wild type.

    IL-20-mediated filaggrin reduction is STAT3-independent

    Because signalling through the type 1 IL-20 receptor phosphorylates STAT3,16 we hypothesised that STAT3 activation is responsible for the IL-20 subfamily-mediated decrease of FLG and FLG2 in squamous epithelial cells. To verify our hypothesis, we stimulated the oesophagus squamous cell carcinoma cell line KYSE-180 with IL-19, IL-20 and IL-24 confirming that the IL-20 subfamily cytokines activate STAT3 (online supplemental figure 6A and online supplemental figure 8A). Interestingly, stimulation with IL-19, IL-20 and IL-24 also resulted in ERK1/2 phosphorylation, while NF-κB (p65) was not activated (online supplemental figure 6B,C, and online supplemental figure 8B,C). Surprisingly, subsequent treatment of IL-20R-expressing KYSE-180 cells (online supplemental figure 6D) with IL-20 subfamily cytokines and the STAT3 inhibitor cucurtabicin 1 resulted in a further decrease of FLG, FLG2 and SPINK7 expression (online supplemental figure 6E). Accordingly, cucurtabicin 1 also lowered FLG, FLG2 and SPINK7 in patient-derived oesophageal organoids stimulated with IL-19, IL-20 and IL-24 (online supplemental figure 6F). To corroborate the impact of STAT3 signalling on the oesophageal epithelium, we crossed Stat3flox/flox with Krt5-CreERT2 mice to generate animals with a tamoxifen-inducible, squamous epithelium-specific deletion of STAT3. Immunofluorescence confirmed the deletion of epithelial STAT3 on tamoxifen injection (online supplemental figure 6G). Tamoxifen-injected Stat3ΔKrt5 mice have increased eosinophils and SiglecF-CD45+ immune cells in the oesophagus compared with tamoxifen-treated Stat3flox/flox littermates demonstrated by H&E staining (figure 6A,B) and flow cytometry analysis (figure 6C–E and online supplemental figure 6H,I), with highest numbers in sens+chal mice. In contrast, Flg2 expression in the oesophagus of Stat3ΔKrt5 mice was drastically reduced in non-chal and sens+chal animals (figure 6F,G). These data reveal that the specific deletion of STAT3 signalling in the squamous epithelium increases epithelial inflammation and interferes with IL-20 subfamily-mediated filaggrin downregulation.

    Figure 6

    STAT3 deletion in oesophageal epithelial cells aggravates experimental EoE. (A) H&E staining of oesophageal sections obtained from Stat3flox/flox and Stat3ΔKrt5 animals. Arrows indicate eosinophils. Scale bars, 50 µm. (B) Counted eosinophils (per HPF) from (A). (C–E) Flow cytometry of oesophageal cell isolates from Stat3flox/flox and Stat3ΔKrt5 mice of indicated treatment groups. (C) Representative flow cytometry blots showing CD45+Siglec-F+ eosinophils and Siglec-FCD45+ immune cells in the oesophagus. Numbers indicate the percentage of eosinophils and Siglec-FCD45+ immune cells. Dot blots showing (D) absolute eosinophil and (E) Siglec-FCD45+ immune cell numbers in the oesophagus. (F,G) Flg2 staining of oesophagus sections from Stat3flox/flox and Stat3ΔKrt5 mice of indicated treatment groups. (F) Representative images; scale bars, 50 µm; (G) Percentage of Flg2 stained area analysed with Fiji (ImageJ V.2.0.0-rc-68/1.52 hour). Data are presented as individual values with medians, with each dot representing one biological replicate. *P<0.05, by Mann-Whitney U test. ctrl, non-sensitised+non-challenged; EoE, eosinophilic oesophagitis; HPF, high-power field; non-chal, sensitised+non-challenged; non-sens, non-sensitised+challenged; sens+chal, sensitised+challenged.

    IL-20 subfamily-mediated epithelial barrier impairment is ERK-dependent

    Reduced expression of the filaggrin family in human oesophageal organoids stimulated with IL-19, IL-20 and IL-24 prompted us to investigate whether the IL-20 subfamily regulates epithelial barrier integrity in the oesophagus. We, therefore, established air–liquid interface (ALI) cultures from biopsy-derived primary keratinocytes and stimulated them with IL-19, IL-20 and IL-24. Transepithelial electrical resistance (TEER), reflecting the integrity and permeability of the epithelium, was reduced in ALI cultures treated with IL-20 subfamily cytokines, substantiating the role of this cytokine family in the regulation of epithelial barrier integrity (figure 7A). While adding the STAT3 inhibitor cucurtabicin 1 to the medium aggravated TEER reduction, the addition of the MAPK (ERK1/2) inhibitor PD98059 rescued the IL-20 subfamily-mediated TEER reduction (figure 7A), confirming our previous observations (figure 6 and online supplemental figure 6). Increased macromolecular flux by fluorescein isothiocyanate (FITC)–dextran (4 kDa) verified increased epithelial permeability in patient-derived ALI cultures stimulated with the IL-20 subfamily cytokines, which was prevented by PD98059 treatment (figure 7B). Consistently, cucurtabicin 1 treatment increased the permeability in the FITC–dextran flux assay (figure 7B). Furthermore, H&E staining visualises the lack of a cornified epithelial layer in patient-derived ALI cultures stimulated with IL-20 subfamily cytokines compared with non-stimulated and PD98059-treated ALI cultures, reflecting reduced epithelial differentiation (figure 7C).

    Figure 7

    ERK inhibition restores barrier function and attenuates experimental EoE. (A) TEER and (B) macromolecule flux of ALI cultures from primary human oesophageal epithelial cells (n=4) stimulated with IL-20 subfamily cytokines (IL-19, IL-20 and IL-24), IL-20 subfamily cytokines and the ERK inhibitor PD98059, and IL-20 subfamily cytokines and the STAT3 inhibitor cucurtabicin 1. (C) Representative H&E staining of non-stimulated and with IL-20 subfamily cytokines or IL-20 subfamily cytokines and PD98059-stimulated ALI cultures. Scale bars, 50 µm. (D) Representative dot blots from oesophagus cell isolates measured by flow cytometry of dimethyl sulfoxide (DMSO) vehicle and PD98059-treated ctrl and sens+chal WT animals. (E) Percentage and absolute numbers of CD45+ cells, (F) Siglec-F+CD45+ eosinophils and (G) Siglec-F-CD45+ cells from indicated experimental groups. (H) H&E staining of oesophageal sections of DMSO (vehicle) and PD98059-treated ctrl and sens+chal WT animals. Black arrowheads indicate eosinophils. Scale bars, 50 µm. (I) Quantification of eosinophils (per HPF) from (H). Data are presented as individual values with medians, with each dot representing one biological replicate. Each dot (A,B) represents the mean of four biological replicates; error bars represent SD. *P<0.05, **P<0.01, ***P<0.001, by Mann-Whitney U test or two-way analysis of variance. ALI, air–liquid interface; ctrl, non-sensitised+non-challenged; DMSO, dimethyl sulfoxide; EoE, eosinophilic oesophagitis; HPF, high-power field; IL, interleukin; NS, non-stimulated; sens+chal, sensitised+challenged; TEER, transepithelial resistance; WT, wild type.

    Collectively, these data support a functional impairment of the oesophageal epithelial barrier mediated through IL-20 subfamily cytokine stimulation.

    The observation that inhibition of the ERK pathway may forestall epithelial impairment raised the question of whether ERK inhibition could serve as a therapeutic option in EoE. To answer this question, we treated sens+chal WT mice with PD98059, resulting in decreased numbers of infiltrating eosinophils and SiglecFCD45+ immune cells in the oesophagus compared with vehicle-treated chal+sens mice (figure 7D–G). H&E staining of oesophageal sections confirmed lower eosinophil infiltration in PD98059-treated sens+chal mice (figure 7H,I). These results suggest that ERK inhibition can prevent IL-20 subfamily-mediated impairment of the oesophageal epithelial barrier attenuating experimental EoE in mice.

    Discussion

    In this study, by combining transcriptomics and proteomics of patient-derived oesophageal organoids together with functional assays in patient-derived ALI cultures and animal models, we show that dysregulated IL-20 subfamily cytokine expression in EoE downregulates the expression of genes and proteins involved in epithelial cornification and differentiation including the filaggrin family. Our data suggest that the IL-20 subfamily contributes to epithelial barrier impairment in EoE. Targeting the IL-20 subfamily signalling pathway using MAPK inhibitors might be a potential therapeutic strategy to restore epithelial barrier integrity in EoE.

    We found increased expression of IL-20 subfamily cytokines in individuals with active EoE, which decreased in inactive EoE after treatment with topical corticosteroids. Importantly, elevated IL-19 and IL-20 concentrations have been noted in the oesophagus and the serum. Nevertheless, further investigations are required to assess whether IL-20 subfamily serum concentrations potentially qualify as biomarkers to predict EoE activity. However, the cellular sources of IL-19, IL-20 and IL-24 in the oesophagus have not been explored. Persisting ambiguities concerning the cellular origins of IL-19, IL-20 and IL-24 lead us to conduct experiments with Il19tdT reporter animals, illustrating that macrophages in the oesophagus are one potential source of IL-19. Similar to the skin, fibroblasts and keratinocytes embody different possible sources of IL-20 subfamily cytokines in the oesophagus.34 Because all three cytokines have a typical expression pattern, the genetic development of reporter mouse lines or fate-mapping experiments may help to unravel further sources of IL-20 subfamily members in the oesophagus.

    The interplay between immune cells, the cytokines they produce and the epithelium has been proposed to be a decisive factor in the development of EoE.8 35–37 Genome-wide association studies and candidate-gene analysis identified EoE risk genes to be expressed mainly by the oesophageal epithelium and a substantial part to be involved in maintaining an intact epithelial barrier.9 38 39 Expression of the type 1 IL20R by the oesophageal epithelium implies epithelial cells as the target of IL-20 subfamily cytokines. Therefore, previous findings showing overexpression of IL-20 and IL-24 results in hyperproliferation of keratinocytes and epidermal hyperplasia,15 22 and reports of impaired barrier function in active EoE due to reduced epithelial barrier components5–8 11 40 tempted us to elaborate whether stimulation of patient-derived oesophageal organoids with IL-20 subfamily cytokines may disrupt the integrity of the epithelium. Bulk RNA-seq and mass spectrometry-based proteomics revealed a reduction of transcripts and proteins involved in cornification and epithelial differentiation on stimulation of patient-derived oesophageal organoids with IL-20 subfamily cytokines. Apart from the filaggrin family responsible for keratin-filament alignment and prevention of water loss in the stratum corneum,41 other essential epithelial barrier constituents involved in tight junction, adherens junctions and desmosome formation were downregulated. Although all these factors are likely to play a role in epithelial barrier function, filaggrins were the most extensively regulated candidates on stimulation with the IL-20 subfamily cytokines. The reduction of IL-20 subfamily cytokines and coinciding restoration of filaggrin expression in inactive EoE implies that the IL-20 subfamily-mediated impairment of the epithelial barrier is a reversible consequence of the inflammation and not a constant disease-inherent feature of EoE. Since mutations in the filaggrin family, proposed as EoE risk genes,8 9 result in an inherent increase in epithelial permeability,42 it is tempting to speculate that the increased expression of IL-20 subfamily cytokines is the outcome of inflammation induced by an inherently increased permeability further amplifying the barrier defect in EoE.

    Comparable to human EoE, the experimental EoE mouse model induced expression of IL-20 subfamily cytokines and reduced expression of filaggrins, while the development of experimental EoE and loss of Flg2 was not precluded in Il-19-deficient animals. Disruption of IL-20 subfamily signalling by genetic deletion of IL-20Rβ, required for both the type 1 and type 2 IL-20 receptors, hampered experimental EoE in Il20R2−/− mice and preserved Flg2 expression in the oesophagus. Thus, we concluded that all three IL-20 subfamily members facilitate the development of EoE. Apart from their effects on epithelial cells, members of the IL-20 subfamily may also act on immune cells. Recent work has suggested that IL-19 supports neutrophil development,43 and IL-24 represses the IL-17 cytokine programme in Th17 cells.44 Whether the IL-20 subfamily cytokines elicit effects on neutrophils and T cells promoting the development of EoE remains to be explored. The expression pattern of the IL-20 receptor subunits in murine BMDMs and single-cell RNA-seq of CD45+ cells from mice with EoE, however, questions a direct effect of the IL-20 subfamily on immune cells in the oesophagus through the known heterodimeric receptors.

    Activating the type 1 and type 2 IL-20 receptors by IL-19, IL-20 and IL-24 induces STAT3 phosphorylation.16 Interestingly, the specific deletion of STAT3 in colonic epithelial cells leads to impaired wound healing, hyperproliferation and high colitis susceptibility.45 Conversely, the specific deletion of STAT3 in epithelial cells protected from colitis-associated cancer in mice.46 47 In line with this, our experiments with STAT3 inhibition in vitro or its squamous epithelium-specific deletion in vivo suggested that blocking STAT3 signalling further impaired the epithelial barrier. Considering our data, we cannot conclude whether JAK/STAT signalling inhibition, successfully used to treat IBD,48 49 also represents an option for the treatment of EoE. However, it must be considered that JAK inhibitors will also constrain STAT activation in immune cells and thereby the production of cytokines facilitating EoE.50

    However, our investigations suggest that STAT3 activation in conserved signalling cascades through the type 1 and type 2 IL-20 receptors does not explain the impairment of epithelial integrity after stimulation with IL-20 subfamily cytokines.

    On the other hand, pharmacological inhibition of ERK1/2 prevented an impaired barrier function in ALI cultures stimulated with IL-19, IL-20 and IL-24, implying IL-20 subfamily-mediated barrier impairment to be ERK1/2-dependent. Comparable observations have been made in the skin, where the treatment of keratinocytes with the specific ERK1/2 inhibitor PD98059 overturned IL-17-mediated filaggrin downregulation.51 Moreover, the IL-1 family cytokine IL-33 also decreases dermal filaggrin expression in an ERK1/2-dependent manner.52 Although we cannot exclude the involvement of other factors in the ERK1/2-dependent regulation of epithelial permeability, IL-20 subfamily signalling seems to display an essential element for epithelial barrier function in the oesophagus. In particular, because alongside IL-20 subfamily-specific downregulation of the GO categories ‘tight junction’, ‘desmosome’, ‘epidermis morphogenesis’ and ‘apical junction complex’, the comparison of our bulk RNA-seq dataset with datasets of IL-13 stimulated oesophageal keratinocytes,6 SPINK7-deficient oesophageal keratinocytes13 and the EoE transcriptome53 unveiled overlapping downregulation of genes involved in epidermis development. These results further augment the image of coinciding mechanisms regulating epithelial barrier function in EoE. IL-13 has been postulated to be a critical molecular driver of EoE, recapitulating a majority of the EoE transcriptome and similarly inducing epithelial barrier impairment.5–7 Intriguingly, patient-derived organoids stimulated with IL-20 subfamily cytokines upregulated the IL-13 receptor subunits, IL-4Ra, IL-13Ra1 and IL-13Ra2. This suggests that the IL-20 subfamily increases epithelial sensitivity for IL-13, reinforcing IL-13-mediated changes in EoE. SPINK7, a member of the SPINK-type family, is highly enriched in the oesophageal epithelium and emerged as an essential regulator of epithelial barrier homeostasis controlling the proteolytic activity of kallikrein 5 and other proteases.13 14 Reduced SPINK7 expression in the epithelium of patients with EoE leads to protease–protease inhibitor imbalance resulting in degradation of epithelial barrier proteins and production of proinflammatory cytokines.13 14 Although IL-13 broadly regulates oesophagus-specific genes,5 6 SPINK7 is not regulated by IL-13.13 Remarkably, SPINK7 was downregulated in patient-derived organoids stimulated with IL-20 subfamily cytokines. Therefore, we suggest specific IL-20 subfamily-mediated effects on the oesophageal epithelium, which position the IL-20 subfamily upstream of SPINK7 and alongside IL-13-mediated changes.

    In summary, our study uses in vitro and in vivo models to conceivably establish an ERK1/2-dependent relationship between the IL-20 subfamily and the oesophageal epithelium in the context of EoE. However, before specific targeting of IL-19, IL-20 and IL-24 and their signalling pathway can be considered for restoring epithelial integrity in EoE, the complex interplay between immune cells, the IL-20 cytokines and the oesophageal epithelium needs to be consolidated in further translational and clinical studies. Identifying possible interfaces between the immune system and the oesophageal epithelium will be critical to dissecting the immune response responsible for developing EoE and identifying novel targets for treatment.

    Data availability statement

    Data are available in a public, open access repository. Data are available upon reasonable request. Bulk RNA sequencing (GEO accession GSE181261; https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE181261), single-cell RNA-seq (GEO accession GSE190482; https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE190482) and proteomics (PXD031509; project DOI: 10.6019/PXD031509) datasets are publicly available in open access repositories.

    Ethics statements

    Patient consent for publication

    Ethics approval

    After obtaining written informed consent, patients were enrolled in the study. The ethics committee for Northwest and Central Switzerland approved the study protocol (ethics protocol EKBB 2019–00273).

    Acknowledgments

    We thank the Genomics Facility Basel, Department of Environmental Systems Science, ETH Zurich, for helping with the RNA sequencing (RNA-seq) and single-cell RNA-seq and the Proteomics Facility Basel, Biozentrum, University of Basel for helping with the mass spectrometry. We also thank the DBM Histology Core Facility, University of Basel, for the support and execution of H&E and immunohistochemistry staining. We further thank the DBM Microscopy Core Facility, University of Basel, for imaging and QuPath analysis support. Calculations were performed at sciCORE (http://scicore.unibas.ch/) Scientific Computing Center at the University of Basel. The Swiss National Science Foundation (SNSF) grant 310030_175548 to JHN financed the publication of our manuscript without restrictions. The SNSF MD PhD fellowship 323530_183981 supported TK. The SNSF grant 32003B_204751/1 supports the Swiss EoE Cohort Study Group.

    References

    Supplementary materials

    • Supplementary Data

      This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Footnotes

    • Collaborators Swiss EoE Cohort Study Group: Patrick Aepli (Division of Gastroenterology and Hepatology, Kantonsspital Lucerne), Luc Biedermann, Philipp Schreiner, Alex Straumann (Division of Gastroenterology and Hepatology, University Hospital Zurich), Annett Franke (Division of Gastroenterology and Hepatology, Kantonsspital Sankt Gallen), Thomas Greuter, Catherine Saner, Alain M Schoepfer (Division of Gastroenterology and Hepatology, Centre Hospitalier Universitaire Vaudois and University of Lausanne), Pascal Juillerat (Division of Gastroenterology and Hepatology, Inselspital/University Hospital Bern), Peter Netzer (GastroZentrum Netzer, Lindenhofspital, Bern), Jean-Benoit Rossel, Ekaterina Safroneeva (Institute of Social and Preventive Medicine, University of Bern), Dagmar Simon (Division of Dermatology, Inselspital/University Hospital Bern), Hans-Uwe Simon (Institute of Pharmacology, University of Bern).

    • Contributors TK, BK, PW and HM carried out experiments. JR performed bioinformatics analysis of the transcriptomics and proteomics data. FO provided the Il20R2−/− mice. SN supported us in establishing a protocol for generating organoids derived from the oesophagus, and MN helped us set up the experimental eosinophilic oesophagitis (EoE) mouse model. TK, AG, PH and JHN recruited patients with EoE. TK and JHN designed the study and wrote the paper. JHN conceived the idea and is the guarantor of the study. All authors discussed the data and read and approved the manuscript.

    • Funding This study was funded by SNSF (grant numbers 310030_175548, 32003B_204751/1 and 323530_183981).

    • Competing interests None declared.

    • Patient and public involvement Patients and/or the public were not involved in the design, conduct, reporting or dissemination plans of this research.

    • Provenance and peer review Not commissioned; externally peer reviewed.

    • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.